Genetic structure in two northern muriqui populations ( Brachyteles hypoxanthus , Primates , Atelidae ) as inferred from fecal DNA

We assessed the genetic diversity of two northern muriqui (Brachyteles hypoxanthus Primata, Atelidae) populations, the Feliciano Miguel Abdala population (FMA, n = 108) in the Brazilian state of Minas Gerais (19°44’ S, 41°49’ W) and the Santa Maria de Jetibá population (SMJ, n = 18) in the Brazilian state of Espírito Santo (20°01’ S, 40°44’ W). Fecal DNA was isolated and PCR-RFLP analysis used to analyze 2160 bp of mitochondrial DNA, made up of an 820 bp segment of the gene cytochrome c oxidase subunit 2 (cox2, EC 1.9.3.1), an 880 bp segment of the gene cytochrome b (cytb, EC 1.10.2.2) and 460 bp of the hypervariable segment of the mtDNA control region (HVRI). The cox2 and cytb sequences were monomorphic within and between populations whereas the HVRI revealed three different population exclusive haplotypes, one unique to the SMJ population and two, present at similar frequencies, in the FMA population. Overall haplotype diversity (h = 0.609) and nucleotide diversity (π = 0.181) were high but reduced within populations. The populations were genetically structured with a high fixation index (FST = 0.725), possibly due to historical subdivision. These findings have conservation implications because they seem to indicate that the populations are distinct management units.

The muriqui or woolly spider monkey (Brachyteles Spix 1823: Primates, Atelidae), endemic to Brazil, is the largest Neotropical primate and was once widespread in the southeastern Atlantic Forest of Brazil.Aguirre (1971) estimated a total population of 2,791-3,226 muriquis, contrasting with a population of about 400,000 he reckoned would have existed in 1500.In the last decade this genus has been split into two species (Groves, 2005), the northern muriqui (Brachyteles hypoxanthus Kuhl 1820) and the southern muriqui (Brachyteles arachnoides É. Geoffroy 1806).Current population estimates for the northern muriqui have indicated at least 864 individuals in the wild and data available for the southern muriqui suggest a minimum population of about 1,300 (Melo and Dias, 2005).
The IUCN Red List cites the northern muriqui as a critically endangered species because only about 900 individuals are known dispersed in 12 populations, five of which contain less than 20 members.Some of the populations are restricted to small, unprotected and isolated forest fragments extending from the south of the state of Bahia throughout Minas Gerais and Espírito Santo states as well as along the Mantiqueira Mountains (Serra da Mantiqueira) on the borders of the states of Minas Gerais, Rio de Janeiro and São Paulo (Rylands et al., 2003b;Mendes et al., 2005).The southern muriqui is distributed along the Serra do Mar from the south of Paraná state to Rio de Janeiro state, is listed as endangered in the IUCN Red List and no more than 1300 are known to occur in relatively large and well protected areas (Melo and Dias, 2005).
The small size and fragmented distribution of northern muriqui populations compromises the ecological viability of this species (Brito and Grelle, 2006).Their advanced age of 9 years at first reproduction and long inter-birth interval of three years, make the small persistent populations more vulnerable to unfavorable demographic conditions than primates with faster life histories or larger populations (Strier et al., 2006).Habitat reduction and hunting have probably forced surviving northern muriqui populations into bottlenecks and reduced or eliminated opportunities for gene flow through populations, resulting in changes in the frequency of alleles, loss of alleles, or both (Young and Clarke, 2000) similar to those documented in the golden lion tamarin (Grativol et al., 2001).
Population viability analysis (PVA) has suggested that only muriqui populations of at least 700 monkeys would be genetically viable (Strier, 1993(Strier, /1994;;Brito and Grelle, 2006) but no empirical genetic data was included in these analyses.Although a preliminary allozyme analysis comparing one population of each species reported a high fixation index (Pope, 1998), knowledge of genetic structure of more populations is crucial for muriqui conservation planning (Fagundes, 2005).
Standardization of a non-invasive method for DNA extraction from muriqui feces is an important tool for assessing the conservation status and behavioral ecology of these monkeys (Chaves et al., 2006).Furthermore, data on the levels of genetic variability and differentiation in muriqui populations as well as pedigree reconstruction and information regarding the relatedness between individuals and the extent of inbreeding can contribute to the planning of effective conservation strategies for these species.
We used the Polymerase Chain Reaction and Restriction Fragment Length Polymorphism analysis (PCR-RFLP) to evaluate the genetic diversity of two northern muriqui populations and characterize the distribution of genetic variability within and between populations.Fecal samples from 126 free-ranging northern muriquis (Brachyteles hypoxanthus Kuhl 1820) were collected immediately after defecation and stored at 4 °C or room temperature (~24-28 °C) in 50 mL polypropylene vials containing a layer, about 1 mm to 4 mm deep, of desiccated silica beads to dehydrate the feces (Chaves et al. 2006).Each individual sampled was identified by its natural markings by the experienced field researchers who collected the fecal samples within the ambit of a collecting license (number 363/2001) issued by the Brazilian National Environmental Agency (Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renováveis -IBAMA).Two populations, separated by 150 km, were studied (Figure 1).One population is in a private conservation unit (Reserva Particular do Patrimônio Natural Feliciano Miguel Abdala, RPPN-FMA) of 957 hectares (ha) located in the state of Minas Gerais (19°44' S, 41°49' W) and is designated the FMA population.This population is made up of four groups (three mixed-sex and one only-male group) totaling 226 northern muriquis (Strier et al., 2006) of which nearly 50% (n = 108) were sampled between 2001 and 2002, once this population has been monitored over the last 25 years by K. B. Strier and her coworkers and animals are used to the hu-man presence.The other population, designated the SMJ population, is in privately owned forest fragments in the municipality of Santa Maria de Jetibá (SMJ) in the state of Espírito Santo (20°0' S, 40°44' W).The forest patches containing the muriqui range from 60 ha to 350 ha and are highly fragmented due to agricultural activities and unevenly connected to each other by hilltop corridors.This metapopulation has been estimated to be comprised of 115 muriqui (Mendes et al., 2005), although only about 16% (n = 18) could be sampled because monitoring of this population only started in 2001.
The specificity of the HVRI primers was shown by their inability to amplify human DNA, probably due to the 19 mismatches (12 forward and 7 reverse) between the primers, which had been derived for monkey sequences, and human DNA.An in silico restriction simulation was performed with Brachyteles HVRI sequence from Genbank using BIOEDIT 7.0.5.3 (Hall, 1999) to identify restriction sites and restriction fragment sizes.Analysis of the human HVRI restriction sites excluded cross-contamination of muriqui PCRs with human DNA (Figure 2).The PCR was carried out in a final volume of 50 μL of 1X Taq buffer containing 3 mM MgCl 2 , 0.4 mM of each dNTP, 0.4 μM of each primer, 2.5 units of Platinum Taq DNA polymerase (Invitrogen) and 50 ng of DNA (Chaves et al., 2006).Amplifications were conducted at 92 °C for 5 min, followed by 35-37 cycles of 92 °C for 60 s, 47 °C to 52 °C for 30 s to 60 s and 72 °C for 30 s to 60s, with a final extension at 72 °C for 5 min.After amplification the amplicons were electrophoretically sized on 1% (w/p) agarose gel using 1 kb and 100 bp Ladders (Invitrogen).Mock PCR blanks were included to check for contamination.
Estimates of the number of nucleotide substitutions (genetic polymorphisms) took into account the previous in silico analysis shown in Figure 2. Three to seven restriction endonucleases (Table 1) digested roughly 20 ng of the PCR amplicons following the manufacturer's protocol.Restriction fragment sizes were determined using polyacrylamide gel electrophoresis (5 to 10% w/v) and ethidium bromide staining, with the gels being photographed under ultraviolet light.Restriction fragment sizes were estimated using 1 kb and 100 bp ladders as reference (detailed procedure is available upon request).Table 1 summarizes the PCR-RFLP results obtained after cutting PCR amplicons.
For the restriction fragment data analysis, one allele was represented by one fragment obtained with an enzyme, and the set of all alleles of a specific digestion (i.e., the cleavage pattern) received a capital letter (Table 1).The composite haplotype panel for each monkey included the cleavage pattern of all the enzymes (Table 2, see Bates, 2002 for details).Genetic diversity analyses were calculated using ARLEQUIN 3.01 (Excoffier et al., 2005) after converting composite haplotypes into a binary matrix (Table 2) based on the presence or absence of alleles (restriction fragments).The population parameters calculated were haplotype frequency, haplotype diversity (h), nucleotide diversity (π), mean number of pairwise difference between haplotypes, and Wright's fixation index (F ST , Wright, 1951).
All 80 restriction assays for cox2 (n = 49) and cytb (n = 31) resulted in monomorphic haplotypes for the seven endonucleases (Table 2).Therefore, we concentrated on analyzing all 126 muriquis for the HVRI segment based on the following assumptions.Firstly, amplification success for HVRI was 100% (against 20%-30% for cox2 and cytb), which is amongst the highest rate ever reported for fecal DNA (Broquet et al., 2007 and references therein).This is likely to be due to the specificity of the primers we used, as well as the short-length segment, which are better suited to analyses of highly degraded fecal DNA.Secondly, HVRI seems to be one of the most variable segments within mammalian mtDNA (Aquadro and Greenberg, 1983;Sbisà et al., 1997), and is thus highly suitable for revealing withinpopulation variation.Finally, mtDNA evolves as a single locus (Avise, 2004), thus analyses of the most variable segment can provide a rough estimate of the diversity of the entire mtDNA genome.
The HVRI assay revealed 35 restriction fragments (Table 2), of which 17 (48.6%)were polymorphic.The MseI and RsaI restriction enzymes generated more fragments (n = 8) and BsaI generated fewer fragments (n = 2), while HinfI generated the longest fragment (460 bp, with no restriction site).Based on the restriction site map shown in Figure 2 one polymorphic restriction site was present for HinfI (position 240 or 220), two sites were present for RsaI (positions 56 and 313), and one site each was present for MseI (position 251) and Tsp509I (around position 400 to 420).The NlaIII and HpyCH4IV sites were invariable.After assembling the single digestion patterns generated by each endonuclease the HVRI, RFLP identified three composite haplotypes, with haplotype SMJ1 being exclusively present in all the 18 monkeys from the SMJ group, while the other two haplotypes (FMA1 and FMA2) were unique to the FMA population and were nearly equal in frequency at 57.4% for FMA1 and 42.6% for FMA2 (Table 2).Differences in pairwise distances between haplotypes (SMJ1/ FMA1 = 12, SMJ1/FMA2 = 13 and FMA1/FMA2 = 9) demonstrated the closer relationship between FMA1 and FMA2 than between either of these haplotypes and the 168 Genetic structure of northern muriqui Figure 2 -The in silico restriction analysis resulted in the haplotype A?AABAA in the muriqui sequence, which is highly similar to muriqui haplotypes found in the SMJ and FMA populations.However, this haplotype was not in our assessment, possibly because this sequence is from an individual from a different population.The human haplotype is markedly different from all muriqui haplotypes, indicating that if human DNA were amplified due to inadvertent contamination it would be readily identified.The BsaI and HpyCH4IV restriction enzymes cannot cut the human segment (GenBank AC_000021/Homo sapiens), while the HinfI restriction enzyme cannot cut muriqui DNA.
SMJ1 haplotype.This finding suggests that the SMJ and FMA populations have been isolated long enough to prevent gene flow and haplotype sharing.
Overall haplotype (h = 0.609 ± 0.022) and nucleotide diversity (π = 0.181 ± 0.095) were relatively high.Withinpopulation diversity indices were slightly lower in the FMA population (h = 0.494 ± 0.016; π = 0.127 ± 0.070) than in the SMJ population, which were both zero.The low number of haplotypes homogeneously distributed is an unfavorable scenario for the genetic diversity of the FMA population, but it is better than the single haplotype found in the SMJ population.
We also found a high fixation index of F ST = 0.725 (p < 0.001), which is strongly indicative of genetic distinctiveness between the SMJ and FMA populations.This F ST is considerably higher than the F ST = 0.413 for allozyme polymorphisms of 12 muriqui (two from a what is now considered a B. arachnoides population and ten from a B. hypoxanthus population different to that studied by us) reported by Pope (1998), who used this data to recommended the elevation of the southern and northern forms of the muriqui into separate species (B.arachnoides and Brachyteles hypoxanthus respectively).
Populations containing mtDNA haplotypes at significantly different frequencies or, in more extreme cases, pre-  -These fragments (15 bp and 25 bp) were not visualized in the gels, probably because they were lost.
senting population-exclusive haplotypes, have been referred as management units and seem to have been historically separated during evolutionary time (Moritz, 1994).
In our study, since the FMA and SMJ northern muriqui populations did not share haplotypes, it is reasonable to assume that they should be given management unit status and actively managed.In addition, the FMA and SMJ populations are connected by low levels of gene flow, are functionally independent and carriers of a portion of the species evolutionary legacy.As a caveat, caution must be exercised with this conclusion since broadening our sampling could reveal a different scenario with admixture of haplotypes.High genetic diversity and F ST seem to be intrinsic to Brachyteles, and may be an important factor contributing to the persistence of small, isolated populations.Relatively few genetic studies involving fecal DNA have been conducted on New World primates as compared to Old World primates (Oklander et al., 2004).Nonetheless, as our results demonstrate, noninvasive genetic studies of wild muriqui populations can provide important insights for conservation.
Table 2 -Composite haplotypes obtained after PCR-RFLP analysis of HVRI in 126 northern muriqui monkeys from Santa Maria de Jetibá in the Brazilian state of Espírito Santo (SMJ population, n = 18) and Feliciano Miguel Abdala in the Brazilian state of Minas Gerais (FMA population, n = 108).FMA1 and FMA2 represent the two haplotypes from FMA population and SMJ1 the one from SMJ population.The binary code should be read from left to right, a '1' indicating the production of a restriction fragment of the specified size (in bp) and a '0' indicating that the fragment was not detected (see footnote for key).

Haplotype
Composite

Figure 1 -
Figure 1 -Map depicting the easternmost portion of Brazil.Sampling sites of the Brachyteles hypoxanthus populations are highlighted.The FMA population in Minas Gerais (MG) state is roughly 150 km from the municipality of Santa Maria de Jetibá (SMJ) in Espírito Santo (ES) state.

Table 1 -
Restriction enzyme cleavage patterns of the genes cytochrome c oxidase subunit 2 (cox2, 820 bp), cytochrome b (cyt b, 880 bp) and the hypervariable segment of mtDNA control region (HVRI, 460 bp) of the Brachyteles hypoxanthus populations from Santa Maria de Jetibá in the Brazilian state of Espírito Santo (SMJ population, n = 18) and Feliciano Miguel Abdala in the Brazilian state of Minas Gerais (FMA population, n = 108).The total number (n) of muriquis analyzed for each mtDNA segment is shown in parentheses in the first column.